Self-adaptive energy management system and method for hybrid power forklift AGV

Abstract

The invention relates to the technical field of energy management, and discloses a self-adaptive energy management system and method for a hybrid power forklift AGV. Firstly, an initial operation parameter data set is obtained, after data preprocessing is carried out, an improved clustering-proximity algorithm is used for recognizing working conditions, and working condition labels are generated; secondly, establishing a target function based on a self-adaptive energy distribution strategy, solving the target function by using a travel mouse optimization algorithm, outputting optimized operation parameters of the hybrid power forklift AGV, predicting the remaining electric quantity of a power battery, outputting an electric quantity predicted value, establishing an energy fuzzy control model in combination with a working condition label, and outputting a self-adaptive electric distribution strategy; and finally, energy distribution is carried out on the hybrid power forklift AGV, a self-adaptive energy distribution strategy is generated, and self-adaptive energy management is achieved. By processing and analyzing the operation parameter data, the purpose of self-adaptive energy management is achieved, and the method is accurate and objective.

Description

Technical Field

[0001] The present invention relates to the technical field of energy management, and specifically to an adaptive energy management system and method for a hybrid forklift AGV. Background Technique

[0002] Chinese Patent CN118297367B discloses an energy management method, medium and device for a hydrogen-electric hybrid energy storage system. The method specifically includes establishing a hydrogen-electric hybrid energy storage system, setting device performance constraints and power purchase upper limit constraints according to wind-solar power generation and load data; setting a set of energy storage system configuration schemes and an energy optimization unit, outputting the energy storage system configuration scheme to the energy optimization unit to obtain a multi-objective model; taking the minimum value of the sum of external power purchase costs and external hydrogen purchase costs as the objective function to obtain an energy management scheme, and outputting a Pareto optimal solution set according to the multi-objective model and the energy storage system configuration scheme; using the minimum data of the multi-objective model as the ideal point, and finding the solution close to the ideal point in the Pareto optimal solution set as the final configuration scheme to complete energy management. However, when solving the objective function, this invention does not dynamically coordinate the objective weights and real-time solution methods, resulting in insufficient practicability.

[0003] Traditional energy management methods usually lead to a decrease in the battery cycle life due to frequent start-stop, or there are deficiencies in energy management due to low energy utilization rate; at the same time, when performing energy management on a hybrid forklift, due to the energy distribution strategy being unable to adapt to complex working conditions, it is unable to maintain the hybrid drive efficiency when cooperating with a roller for production work, and no optimization algorithm is used to dynamically distribute energy, often resulting in problems such as uneven energy distribution and incoordination of the hybrid system. Summary of the Invention

[0004] In view of the problems in the related art, the present invention provides an adaptive energy management system and method for a hybrid forklift AGV to overcome the above-mentioned technical problems existing in the existing related technologies.

[0005] To solve the above technical problems, the present invention is realized through the following technical solutions:

[0006] The present invention provides an adaptive energy management method for a hybrid forklift AGV, including the following steps:

[0007] S1. Collect the operating parameters of the hybrid forklift AGV to obtain an initial operating parameter data set, perform data preprocessing on the initial operating parameter data set, and then use an improved clustering-proximity algorithm to identify the working conditions, obtain a clustering result, and generate a working condition label;

[0008] S2. Based on the adaptive energy allocation strategy, establish an objective function according to the set of operation parameter data, solve the objective function using the lemming optimization algorithm, find the optimal fitness function value, output the optimized operation parameters of the hybrid forklift AGV, and achieve preliminary adaptive energy management;

[0009] S3. Combine the optimized operation parameters of the hybrid forklift AGV to predict the remaining power of the power battery, output the power prediction value, and then establish an energy fuzzy control model in combination with the working condition label to output the adaptive electric distribution strategy;

[0010] S4. Generate an adaptive energy allocation strategy according to the optimized operation parameters and the adaptive electric distribution strategy of the hybrid forklift AGV, perform energy allocation on the hybrid forklift AGV, and achieve adaptive energy management.

[0011] The invention obtains the initial set of operation parameter data, then performs data preprocessing, uses the improved clustering-proximity algorithm to identify the working conditions, and generates working condition labels; after data preprocessing, the reliability of the data is increased, the working condition identification error is reduced, and the improved clustering-proximity algorithm overcomes problems such as large data classification errors in traditional algorithms. By clustering first and then classifying, it has a more accurate classification effect and greatly reduces the execution time; secondly, an objective function is established based on the adaptive energy allocation strategy, and the lemming optimization algorithm is used to solve the objective function to obtain the optimized operation parameters of the hybrid forklift AGV; this algorithm simulates different behaviors of lemmings in nature, searches for the optimal solution in the search space, uses dynamic balance exploration and development to avoid premature convergence and local optimum, has a fast convergence speed and high stability, ensures that it can adapt to complex working conditions during hybrid drive, and significantly improves the energy utilization efficiency and battery cycle life; then, based on the BP neural network, the remaining power of the power battery is predicted, and an energy fuzzy control model is established in combination with the working condition label to output the adaptive electric distribution strategy; using the power prediction value can predict the threshold power situation in advance, which is convenient for adjusting the energy allocation strategy. Fuzzy control is suitable for complex energy management systems, can better perform energy allocation, reduce energy loss, and ensure the utilization efficiency of the motor; finally, an adaptive energy allocation strategy is generated, energy allocation is performed on the hybrid forklift AGV, and adaptive energy management is achieved, ensuring the power coordination of the hybrid system.

[0012] Preferably, the S1 includes the following steps:

[0013] S11. During the operation of the hybrid forklift AGV, collect the operation parameters of the hybrid forklift AGV, where the operation parameters include driving acceleration, driving speed, engine power, etc., obtain the operation parameter data, and form the initial set of operation parameter data A = {A1, A2, A3,..., A m}, where A mIt represents the initial operating parameter data subset composed of the m-th operating parameter, and each initial operating parameter data subset contains n operating parameter data; and obtain the operating parameter acquisition time point to generate the operating parameter time series where a n represents the n-th operating parameter acquisition time point. Perform data preprocessing on the initial operating parameter data set to obtain a processed operating parameter data set. The specific steps are as follows:

[0014] S111. Select any initial operating parameter data subset in the initial operating parameter data set, denoted as the operating parameter data set to be processed. Set a time window and a change threshold according to the operating parameter time series. Place the time window in the operating parameter data set to be processed, and calculate the change rate of the operating parameter data in the time window. When the change rate of adjacent operating parameter data is greater than the change threshold, regard the operating parameter data corresponding to the change rate of adjacent operating parameter data as suspicious operating parameter data and mark it;

[0015] S112. Move the time window, and calculate the change rate of the operating parameter data in the time window again to obtain the change rate of the suspicious operating parameter data. When the change rate of the suspicious operating parameter data is greater than the change threshold, delete the suspicious operating parameter data at this time, otherwise clear the mark; continue to move the time window until the operating parameter data set to be processed is traversed, and continue to preprocess the initial operating parameter data set to obtain a processed operating parameter data set;

[0016] S12. According to the processed operating parameter data set, combine the improved clustering algorithm and the proximity algorithm to obtain the improved clustering-proximity algorithm. Use the improved clustering-proximity algorithm to identify the working conditions, obtain the clustering results, and generate working condition labels. The specific steps are as follows:

[0017] S121. Select the operating parameter data corresponding to the operating conditions from the processed set of operating parameter data. The operating conditions are the driving acceleration and the driving speed, and form a driving acceleration data set and a driving speed data set. Select p sample data from the driving acceleration data set and the driving speed data set, calculate the distances between the sample data and other data, and select the k data with the farthest distances as the first data cluster class and the second data cluster class respectively. Remove the p sample data from the driving acceleration data set and the driving speed data set to form a new driving acceleration data set and a new driving speed data set, and then calculate the distances between the new driving acceleration data set, the new driving speed data set, the first data cluster class, and the second data cluster class. Select the distance modes as the center points of the first data cluster class and the second data cluster class respectively, and repeat the iteration of the center points of the data cluster classes until the center points of the data cluster classes no longer change, and obtain the first final data cluster class and the second final data cluster class respectively.

[0018] S122. According to the first final data cluster class and the second final data cluster class, divide the driving acceleration data set and the driving speed data set into several driving acceleration data subsets and driving speed data subsets respectively to obtain the clustering result. Add operating condition labels to the driving acceleration data subsets and the driving speed data subsets. When the driving acceleration data in the driving acceleration data subset is 0 and the driving speed data in the driving speed data subset remains unchanged, the operating condition label is light load on flat road at this time. When the driving acceleration data in the driving acceleration data subset is 0 and the driving speed data in the driving speed data subset is 0, the operating condition label is loading and unloading at this time. When the driving acceleration data in the driving acceleration data subset is negative and the driving speed data in the driving speed data subset decreases, the operating condition label is decelerating downhill at this time. When the driving acceleration data in the driving acceleration data subset is positive and the driving speed data in the driving speed data subset decreases, the operating condition label is heavy load climbing uphill at this time, and generate the operating condition labels.

[0019] The present invention preprocesses the initial operating parameter data set, increases the reliability of the data, reduces the operating condition recognition error, and then uses the improved clustering-proximity algorithm to identify the operating conditions and generate the operating condition labels, overcoming problems such as large data classification errors in traditional algorithms. By clustering first and then classifying, it has a more accurate classification effect and greatly reduces the execution time.

[0020] Preferably, the S2 includes the following steps:

[0021] S21. Obtain the bus voltage, battery current, and engine power based on the processed operating parameter data set. Calculate the engine fuel consumption based on the engine power, calculate the battery current change rate based on the battery current. Set δ1 and δ2 to represent dynamic weights, and establish an objective function minf = δ1·B1 + δ2·B2 based on the adaptive energy distribution strategy, where B1 represents the engine fuel consumption and B2 represents the battery current change rate; set the rated voltage and the voltage threshold ω, then where U represents the bus voltage; set the upper and lower bounds of the battery charge. The remaining battery charge of the power battery is greater than or equal to the lower bound of the charge and less than or equal to the upper bound of the charge to obtain the constraint conditions;

[0022] S22. Generate a fitness function according to the objective function and the constraint conditions Use the lemming optimization algorithm to solve the objective function, find the optimal fitness function value, and output the optimized operating parameters of the hybrid forklift AGV. The specific steps are as follows:

[0023] S221. Consider the process of solving the objective function as a search space. Randomly generate a lemming population in the search space. The dimension of the lemming population is j. Initialize the lemming population. The lemming individuals in the lemming population represent candidate solutions. The candidate solutions include the engine power, the bus voltage compensation amount, and the battery current; the lemming population enters the exploration stage. Set the probability density function of the normal distribution with a variance of 1 and a mean of 0 to generate the Brownian motion step size, and obtain the random number vector β of the Brownian motion. c represents a vector of size 1×j and the vector elements are between the interval [-1, 1]. φ represents the search direction parameter; set the current iteration number to t. The position of the rth lemming individual in the lemming population is X r (t), the position of the ith lemming individual in the lemming population is X i (t), the position of the best lemming individual at the tth iteration is X′(t). Update the position X i (t) to obtain the position X i (t + 1) of the ith lemming individual in the lemming population at the (t + 1)th iteration. X i (t + 1) = X′(t) + φ·β·(c·(X′(t) - X i (t)) + (1 - c)·(X r (t) - X The position of the sth lemming individual in the lemming population is X s (t). Update the position X i (t + 1) to obtain X i (t + 1) = X i (t) + φ·χ·(X′(t) - Xs (t)), and the exploration stage is completed at this time;

[0024] S222. The lemming population enters the development stage. The lemming population adopts a spiral search strategy, calculates the average distance between the current best lemming individual position and the positions of other lemming individuals, which is denoted as the search radius R, and sets b2 to represent a random number between the interval [0, 1]. The spiral coefficient Uses the spiral coefficient to update the position X i (t + 1) to obtain The lemming population evades predators, introduces Lévy flight, denotes the Lévy flight function as Levy(j), sets the maximum number of iterations as T, and the evasion coefficient Updates the position X i (t + 1) again to complete the development stage and obtain X i (t + 1) = X′(t) + φ·γ·Levy(j)·(X′(t) - X i (t)), and generates the next generation of lemming population;

[0025] S223. Uses the energy factor to balance the exploration stage and the development stage. Sets b3 to represent a random number between the interval [0, 1]. Then the energy factor Sets an energy threshold. When the energy factor is greater than the energy threshold, the lemming population enters the exploration stage at this time; otherwise, the lemming population enters the development stage. Continuously iterate until the current number of iterations reaches the maximum number of iterations, then stop iterating to obtain the final lemming population; find the best lemming individual corresponding to the best fitness function value in the final lemming population. The best lemming individual is the global optimal solution, and the global optimal solution includes the optimized engine power, the optimized bus voltage compensation amount, and the optimized battery current, and generates the optimized operating parameters of the hybrid forklift AGV;

[0026] S23. The hybrid forklift AGV operates according to the optimized operating parameters of the hybrid forklift AGV. At this time, the hybrid forklift AGV realizes the preliminary management of adaptive energy.

[0027] The present invention establishes an objective function based on an adaptive energy distribution strategy, uses the lemming optimization algorithm to solve and obtain the optimized operating parameters of the hybrid forklift AGV. This algorithm searches for the optimal solution in the search space by simulating the different behaviors of lemmings in nature, uses dynamic balance of exploration and development to avoid premature convergence and local optimum, has a fast convergence speed and high stability, ensures that it can adapt to complex working conditions during hybrid drive, and significantly improves the energy utilization efficiency and battery cycle life.

[0028] Preferably, the S3 includes the following steps:

[0029] S31. Obtain the optimized engine power, optimized bus voltage compensation amount, and optimized battery current based on the optimized operating parameters of the hybrid forklift AGV, calculate the optimized battery voltage, measure the battery temperature, and generate an optimized operating parameter data matrix with a specification of 3×h in combination with the optimized battery current, optimized battery voltage, and battery temperature, where h represents the number of optimized operating parameter data; collect the battery current, battery voltage, and battery temperature again to form a sample data matrix, then train the BP neural network to obtain the BP neural network prediction model and output the predicted power value. The specific steps are as follows:

[0030] S311. Set the input layer nodes, hidden layer nodes, and output layer nodes of the BP neural network, with the learning rate being λ. Normalize each row of the sample data matrix to obtain the normalized sample data matrix. Divide the normalized sample data matrix into a sample training set and a sample test set. Input the sample training set into the BP neural network for training, and continuously iterate until the BP neural network converges to obtain the trained BP neural network;

[0031] S312. Input the sample test set into the trained BP neural network to output the prediction result. Set the error threshold. When the prediction result error is less than the error threshold, stop training, and at this time, obtain the BP neural network prediction model; otherwise, adjust the weights until the prediction result error is less than the error threshold; Normalize the optimized operating parameter data matrix and input it into the BP neural network prediction model to output the predicted power value;

[0032] S32. Establish an energy fuzzy control model in combination with the predicted power value and the working condition label, and output an adaptive electric distribution strategy. The specific steps are as follows:

[0033] S321. Establish a rule base based on the predicted power value and the working condition label. When the predicted power value is less than the lower power limit, at this time, under all working condition labels, the engine drives alone, and the hybrid forklift AGV is in the charging state to obtain the first rule base; when the predicted power value is greater than or equal to the lower power limit and less than or equal to the upper power limit, at this time, under all working condition labels, the hybrid forklift AGV is in combined drive to obtain the second rule base; when the predicted power value is greater than the upper power limit, at this time, under all working condition labels, the hybrid forklift AGV is in motor drive alone to obtain the third rule base;

[0034] S322. Set the membership function. The membership function is a normal distribution function. The motor power includes g1 fuzzy subsets, the driving speed includes g2 fuzzy subsets, and the predicted power value includes g3 fuzzy subsets. Combine the rule base and the membership function to establish a fuzzy control rule table. According to the change of the driving speed under different working condition labels, input the predicted power value and then output the motor power to complete the energy distribution and obtain the adaptive electric distribution strategy.

[0035] The invention predicts the remaining power of the power battery based on the BP neural network, advances the threshold power condition, facilitates the adjustment of the energy distribution strategy, then establishes an energy fuzzy control model, outputs the adaptive electric distribution strategy. Fuzzy control is applicable to complex energy management systems, can better perform energy distribution, reduce energy loss, and ensure the utilization efficiency of the motor.

[0036] Preferably, the S4 includes the following steps:

[0037] S41. When the hybrid forklift AGV is in combined drive, the hybrid forklift AGV operates according to the optimized operating parameters of the hybrid forklift AGV. When the hybrid forklift AGV is in motor-only drive, the hybrid forklift AGV operates according to the adaptive electric distribution strategy to obtain the adaptive energy distribution strategy, and perform energy distribution on the hybrid forklift AGV during the operation of the hybrid forklift AGV to achieve adaptive energy management.

[0038] This embodiment also discloses a system for an adaptive energy management method for a hybrid forklift AGV, specifically including: a working condition label generation module, an operating parameter optimization module, an electric distribution strategy generation module, and an adaptive energy management module;

[0039] The working condition label generation module is used to perform data preprocessing on the initial operating parameter data and then identify the working conditions to generate working condition labels;

[0040] The operating parameter optimization module is used to solve for the optimized operating parameters using the lemming optimization algorithm;

[0041] The electric distribution strategy generation module is used to establish an energy fuzzy control model and output an adaptive electric distribution strategy;

[0042] The adaptive energy management module is used to generate an adaptive energy distribution strategy according to the optimized operating parameters and the adaptive electric distribution strategy.

[0043] The present invention has the following beneficial effects:

[0044] 1. The invention pre - processes the initial operation parameter data set, which increases the reliability of the data and reduces the working condition recognition error. Then, an improved clustering - proximity algorithm is used to identify the working conditions and generate working condition labels, overcoming problems such as large data classification errors in traditional algorithms. The method of clustering first and then classifying has a more accurate classification effect and greatly reduces the execution time.

[0045] 2. The invention establishes an objective function based on an adaptive energy allocation strategy and uses a lemming optimization algorithm to solve for the optimized operation parameters of the hybrid forklift AGV. By simulating different behaviors of lemmings in nature, this algorithm searches for the optimal solution in the search space, uses dynamic balance exploration and development to avoid premature convergence and local optima, has a fast convergence speed and high stability, ensures adaptability to complex working conditions during hybrid drive, and significantly improves energy utilization efficiency and battery cycle life.

[0046] 3. The invention predicts the remaining power of the power battery based on a BP neural network, advances the threshold power situation to facilitate the adjustment of the energy allocation strategy, and then establishes an energy fuzzy control model to output an adaptive electric allocation strategy. Fuzzy control is applicable to complex energy management systems, can better allocate energy, reduce energy loss, and ensure the utilization efficiency of the motor.

[0047] Of course, it is not necessary for any product implementing the present invention to achieve all the above - mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0049] Figure 1 FIG. is a schematic flow diagram of the adaptive energy management of an adaptive energy management system for a hybrid forklift AGV provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0051] In the description of the present invention, it should be understood that terms such as "open hole", "upper", "lower", "top", "middle", "inner", etc. indicating orientation or positional relationships are only for the convenience of describing the invention and simplifying the description, rather than indicating or implying that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the invention.

[0052] Embodiment 1

[0053] Please refer to Figure 1 , this embodiment discloses an adaptive energy management method for a hybrid forklift AGV, which specifically includes the following contents:

[0054] S1. Collect the operating parameters of the hybrid forklift AGV to obtain an initial set of operating parameter data. Perform data preprocessing on the initial set of operating parameter data, and then use an improved clustering-proximity algorithm to identify the working conditions, obtain the clustering results, and generate working condition labels;

[0055] The S1 includes the following steps:

[0056] S11. During the operation of the hybrid forklift AGV, collect the operating parameters of the hybrid forklift AGV. The operating parameters include driving acceleration, driving speed, engine power, etc., to obtain operating parameter data, and form an initial set of operating parameter data A = {A1, A2, A3,..., A m}, where A m represents the initial set of operating parameter data subsets composed of the mth operating parameter, and each initial set of operating parameter data subsets contains n operating parameter data; and obtain the operating parameter collection time points to generate an operating parameter time series where a n represents the nth operating parameter collection time point. Perform data preprocessing on the initial set of operating parameter data to obtain a processed set of operating parameter data. The specific steps are as follows:

[0057] S111. Select any initial set of operating parameter data subsets in the initial set of operating parameter data, denoted as the set of operating parameter data to be processed. Set a time window and a change threshold according to the operating parameter time series. Place the time window in the set of operating parameter data to be processed, calculate the change rate of the operating parameter data in the time window. When the change rate of adjacent operating parameter data is greater than the change threshold, regard the operating parameter data corresponding to the change rate of adjacent operating parameter data as suspicious operating parameter data and mark it;

[0058] S112. Move the time window, recalculate the change rate of the operating parameter data in the time window to obtain the change rate of the suspicious operating parameter data. When the change rate of the suspicious operating parameter data is greater than the change threshold, delete the suspicious operating parameter data at this time; otherwise, clear the flag. Continue to move the time window until all the operating parameter data sets to be processed are traversed, and continue to preprocess the initial operating parameter data set to obtain the processed operating parameter data set;

[0059] S12. Based on the processed operating parameter data set, combine the improved clustering algorithm and the proximity algorithm to obtain the improved clustering-proximity algorithm. Use the improved clustering-proximity algorithm to identify the working conditions, obtain the clustering results, and generate working condition labels. The specific steps are as follows:

[0060] S121. Select the operating parameter data corresponding to the working condition operating parameters in the processed operating parameter data set. The working condition operating parameters are the driving acceleration and the driving speed, and form a driving acceleration data set and a driving speed data set. Select p sample data in the driving acceleration data set and the driving speed data set, calculate the distances between the sample data and other data, and select the k data with the farthest distances as the first data cluster class and the second data cluster class respectively. Remove the p sample data from the driving acceleration data set and the driving speed data set to form a new driving acceleration data set and a new driving speed data set, and then calculate the distances between the new driving acceleration data set and the new driving speed data set and the first data cluster class and the second data cluster class respectively. Select the distance mode as the center point of the first data cluster class and the center point of the second data cluster class respectively, and repeat the iteration of the data cluster class center points until the data cluster class center points no longer change, and obtain the first final data cluster class and the second final data cluster class respectively;

[0061] S122. Based on the first final data cluster class and the second final data cluster class, divide the driving acceleration data set and the driving speed data set into several driving acceleration data subsets and driving speed data subsets respectively to obtain the clustering results. Add working condition labels to the driving acceleration data subsets and the driving speed data subsets. When the driving acceleration data in the driving acceleration data subset is 0 and the driving speed data in the driving speed data subset remains unchanged, the working condition label is light load on flat road at this time; when the driving acceleration data in the driving acceleration data subset is 0 and the driving speed data in the driving speed data subset is 0, the working condition label is loading and unloading at this time; when the driving acceleration data in the driving acceleration data subset is negative and the driving speed data in the driving speed data subset decreases, the working condition label is decelerating downhill at this time; when the driving acceleration data in the driving acceleration data subset is positive and the driving speed data in the driving speed data subset decreases, the working condition label is heavy load climbing uphill at this time, and generate the working condition labels;

[0062] S2. Based on the adaptive energy allocation strategy, establish an objective function according to the set of operating parameter data, solve the objective function using the lemming optimization algorithm, find the optimal fitness function value, and output the optimized operating parameters of the hybrid forklift AGV to achieve preliminary adaptive energy management;

[0063] S2 includes the following steps:

[0064] S21. According to the processed set of operating parameter data, obtain the bus voltage, battery current, and engine power. Calculate the engine fuel consumption based on the engine power, calculate the battery current change rate based on the battery current. Set δ1 and δ2 to represent dynamic weights, and establish an objective function minf = δ1·B1 + δ2·B2 based on the adaptive energy allocation strategy, where B1 represents the engine fuel consumption and B2 represents the battery current change rate; Set the rated voltage and the voltage threshold ω, then where U represents the bus voltage; Set the upper and lower bounds of the battery charge. The remaining battery charge of the power battery is greater than or equal to the lower bound of the charge and less than or equal to the upper bound of the charge to obtain the constraint conditions;

[0065] S22. Generate a fitness function according to the objective function and the constraint conditions Solve the objective function using the lemming optimization algorithm, find the optimal fitness function value, and output the optimized operating parameters of the hybrid forklift AGV. The specific steps are as follows:

[0066] S221. Regard the process of solving the objective function as a search space. Randomly generate a lemming population in the search space. The dimension of the lemming population is j. Initialize the lemming population. The lemming individuals in the lemming population represent candidate solutions. The candidate solutions include engine power, bus voltage compensation, and battery current; The lemming population enters the exploration stage. Set the probability density function of a normal distribution with a variance of 1 and a mean of 0 to generate a Brownian motion step size, and obtain a random number vector β of Brownian motion. c represents a vector of size 1×j and the vector elements are between the interval [-1, 1]. φ represents the search direction parameter; Set the current iteration number as t. The position of the r-th lemming individual in the lemming population is X r (t), the position of the i-th lemming individual in the lemming population is X i (t), the position of the best lemming individual at the t-th iteration is X′(t). Update the position X i (t) to obtain the position X i (t + 1) of the i-th lemming individual in the lemming population at the (t + 1)-th iteration, X i (t + 1) = X′(t) + φ·β·(c·(X′(t) - X i (t)) + (1 - c)·(X r(t)); The lemming population starts to search for food. Let b1 represent a random number between the interval [0, 1], and the iterative random coefficient The position of the s-th lemming individual in the lemming population is X s (t). For the position X i (t + 1), update it to obtain X i (t + 1) = X i (t) + φ·χ·(X′(t) - X s (t)). At this time, the exploration stage is completed;

[0067] S222. The lemming population enters the exploitation stage. The lemming population adopts a spiral search strategy. Calculate the average distance between the current best lemming individual position and the positions of other lemming individuals, denoted as the search radius R. Let b2 represent a random number between the interval [0, 1], and the spiral coefficient Use the spiral coefficient to update the position X i (t + 1) to obtain The lemming population evades predators. Introduce Levy flight, denote the Levy flight function as Levy(j), set the maximum number of iterations as T, and the evasion coefficient Update the position X i (t + 1) again to complete the exploitation stage and obtain X i (t + 1) = X′(t) + φ·γ·Levy(j)·(X′(t) - X i (t)). Generate the next generation of the lemming population;

[0068] S223. Use the energy factor to balance the exploration stage and the exploitation stage. Let b3 represent a random number between the interval [0, 1], then the energy factor Set the energy threshold. When the energy factor is greater than the energy threshold, at this time the lemming population enters the exploration stage, otherwise the lemming population enters the exploitation stage. Continuously iterate until the current number of iterations reaches the maximum number of iterations, then stop the iteration to obtain the final lemming population; In the final lemming population, find the best lemming individual corresponding to the best fitness function value. The best lemming individual is the global optimal solution. The global optimal solution includes the optimized engine power, the optimized bus voltage compensation amount, and the optimized battery current, and generate the optimized operating parameters of the hybrid forklift AGV;

[0069] S23. The hybrid forklift AGV operates according to the optimized operating parameters of the hybrid forklift AGV. At this time, the hybrid forklift AGV realizes the preliminary self - adaptive energy management;

[0070] S3. Combine the optimized operating parameters of the hybrid forklift AGV to predict the remaining power of the power battery, output the power prediction value, and then establish an energy fuzzy control model based on the working condition label to output an adaptive electric distribution strategy;

[0071] The S3 includes the following steps:

[0072] S31. According to the optimized operating parameters of the hybrid forklift AGV, obtain the optimized engine power, optimized bus voltage compensation amount, and optimized battery current, calculate the optimized battery voltage, measure the battery temperature, and combine the optimized battery current, optimized battery voltage, and battery temperature to generate an optimized operating parameter data matrix with a specification of 3×h, where h represents the number of optimized operating parameter data; collect the battery current, battery voltage, and battery temperature again to form a sample data matrix, and then train the BP neural network to obtain the BP neural network prediction model and output the power prediction value. The specific steps are as follows:

[0073] S311. Set the input layer nodes, hidden layer nodes, and output layer nodes of the BP neural network, with the learning rate being λ. Normalize each row of the sample data matrix to obtain the normalized sample data matrix. Divide the normalized sample data matrix into a sample training set and a sample test set. Input the sample training set into the BP neural network for training, and continuously iterate until the BP neural network converges to obtain the trained BP neural network;

[0074] S312. Input the sample test set into the trained BP neural network, output the prediction result, set the error threshold. When the prediction result error is less than the error threshold, stop training, and at this time, obtain the BP neural network prediction model; otherwise, adjust the weights until the prediction result error is less than the error threshold; normalize the optimized operating parameter data matrix and input it into the BP neural network prediction model to output the power prediction value;

[0075] S32. Combine the power prediction value and the working condition label to establish an energy fuzzy control model and output an adaptive electric distribution strategy. The specific steps are as follows:

[0076] S321. Establish a rule base according to the power prediction value and the working condition label. When the power prediction value is less than the lower power limit, at this time, under all working condition labels, the engine drives alone, and the hybrid forklift AGV is in the charging state, to obtain the first rule base; when the power prediction value is greater than or equal to the lower power limit and less than or equal to the upper power limit, at this time, under all working condition labels, the hybrid forklift AGV is in combined drive, to obtain the second rule base; when the power prediction value is greater than the upper power limit, at this time, under all working condition labels, the hybrid forklift AGV is in motor-alone drive, to obtain the third rule base;

[0077] S322. Set the membership function. The membership function is a normal distribution function. The motor power includes g1 fuzzy subsets, the driving speed includes g2 fuzzy subsets, and the predicted power value includes g3 fuzzy subsets. Combine the rule base and the membership function to establish a fuzzy control rule table. According to the change of the driving speed under different working condition labels, input the predicted power value and output the motor power to complete the energy distribution and obtain an adaptive electric distribution strategy.

[0078] S4. Generate an adaptive energy distribution strategy based on the optimized operating parameters and the adaptive electric distribution strategy of the hybrid forklift AGV, perform energy distribution on the hybrid forklift AGV, and achieve adaptive energy management.

[0079] S4 includes the following steps:

[0080] S41. When the hybrid forklift AGV is in combined drive, the hybrid forklift AGV operates according to the optimized operating parameters of the hybrid forklift AGV. When the hybrid forklift AGV is in motor-only drive, the hybrid forklift AGV operates according to the adaptive electric distribution strategy to obtain an adaptive energy distribution strategy, perform energy distribution on the hybrid forklift AGV during the operation of the hybrid forklift AGV, and achieve adaptive energy management.

[0081] Embodiment 2

[0082] This embodiment also discloses a system for an adaptive energy management method for a hybrid forklift AGV, specifically including: a working condition label generation module, an operating parameter optimization module, an electric distribution strategy generation module, and an adaptive energy management module.

[0083] The working condition label generation module is used to perform data preprocessing on the initial operating parameter data and identify the working conditions to generate working condition labels.

[0084] The operating parameter optimization module is used to solve for the optimized operating parameters using the lemming optimization algorithm.

[0085] The electric distribution strategy generation module is used to establish an energy fuzzy control model and output an adaptive electric distribution strategy.

[0086] The adaptive energy management module is used to generate an adaptive energy distribution strategy based on the optimized operating parameters and the adaptive electric distribution strategy.

[0087] In the description of this specification, the descriptions referring to the terms "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0088] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not exhaust all the details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification in order to better explain the principles and practical applications of the invention, so that those skilled in the art can well understand and utilize the invention.

Claims

1. An adaptive energy management method for a hybrid forklift AGV, characterized in that, It includes the following steps: S1. Collect the operation parameters of the hybrid forklift AGV to obtain the initial operation parameter data set, perform data preprocessing on the initial operation parameter data set, and then use the improved clustering-proximity algorithm to identify the working conditions, obtain the clustering results, and generate working condition labels; S2. Based on the adaptive energy distribution strategy, establish an objective function according to the operation parameter data set, solve the objective function, find the optimal fitness function value, output the optimized operation parameters of the hybrid forklift AGV, and achieve preliminary adaptive energy management; S3. Combine the optimized operation parameters of the hybrid forklift AGV to predict the remaining power of the power battery, output the power prediction value, and then establish an energy fuzzy control model in combination with the working condition labels to output the adaptive electric distribution strategy; S4. Generate an adaptive energy distribution strategy according to the optimized operation parameters and the adaptive electric distribution strategy of the hybrid forklift AGV, perform energy distribution on the hybrid forklift AGV, and achieve adaptive energy management.

2. The intelligent control and power supply method for an Internet of Things-based physical interlock according to claim 1, characterized in that, The S1 includes the following steps: S11. Collect the operation parameters of the hybrid forklift AGV to obtain the operation parameter data, and form the initial operation parameter data set; obtain the operation parameter collection time points, generate the operation parameter time series, and perform data preprocessing on the initial operation parameter data set to obtain the processed operation parameter data set; S12. According to the processed operation parameter data set, combine the improved clustering algorithm and the proximity algorithm to obtain the improved clustering-proximity algorithm, and use the improved clustering-proximity algorithm to identify the working conditions, obtain the clustering results, and generate working condition labels.

3. A method for intelligent control and power supply of an Internet of Things-based physical interlock, according to claim 2, characterized in that The S11 includes the following steps: S111. Select the operation parameter data set to be processed in the initial operation parameter data set, set the time window and change threshold according to the operation parameter time series, calculate the change rate of the operation parameter data in the time window, and mark the suspicious operation parameter data by comparing the change threshold and the change rate; S112. Calculate the change rate of the suspicious operation parameter data. When the change rate of the suspicious operation parameter data is greater than the change threshold, delete the suspicious operation parameter data, otherwise clear the mark; move the time window until all the operation parameter data sets to be processed are traversed to obtain the processed operation parameter data set.

4. A method for intelligent control and power supply of an Internet of Things-based physical interlock, according to claim 3, characterized in that, The S12 includes the following steps: S121. Select the driving acceleration data set and the driving speed data set in the processed operation parameter data set, use the clustering algorithm to determine the first data cluster class and the second data cluster class, and then use the proximity algorithm to repeatedly iterate the center points of the data cluster classes until the center points of the data cluster classes no longer change, and obtain the first final data cluster class and the second final data cluster class respectively; S122. Divide the first final data cluster class and the second final data cluster class to obtain several driving acceleration data subsets and driving speed data subsets, obtain the clustering results, and add working condition labels to the driving acceleration data subsets and the driving speed data subsets.

5. The intelligent control and power supply method for an Internet of Things-based physical interlock according to claim 4, characterized in that, The S2 includes the following steps: S21. Obtain the bus voltage, battery current, and engine power based on the processed set of operating parameter data. Calculate the engine fuel consumption based on the engine power, and calculate the battery current change rate based on the battery current. Set δ1 and δ2 to represent dynamic weights, and establish an objective function minf = δ1·B1 + δ2·B2 based on the adaptive energy distribution strategy, where B1 represents the engine fuel consumption and B2 represents the battery current change rate; set the rated voltage and the voltage threshold ω, then where U represents the bus voltage; set the upper and lower bounds of the battery charge. The remaining battery charge of the power battery is greater than or equal to the lower bound of the charge and less than or equal to the upper bound of the charge to obtain the constraint conditions; S22. Generate a fitness function according to the objective function and the constraint conditions Use the lemming optimization algorithm to solve the objective function, find the optimal fitness function value, and output the optimized operating parameters of the hybrid forklift AGV 6. The intelligent control and power supply method for an Internet of Things-based physical interlock according to claim 5, characterized in that, The S22 includes the following steps: S221. Regard the process of solving the objective function as a search space. Randomly generate a lemming population in the search space. The dimension of the lemming population is j. Initialize the lemming population. The lemming individuals in the lemming population represent candidate solutions. The candidate solutions include engine power, bus voltage compensation amount, and battery current. The lemming population enters the exploration stage. Set the probability density function of the normal distribution with a variance of 1 and a mean of 0 to generate the Brownian motion step size, and obtain the random number vector β of the Brownian motion. c represents a vector of size 1×j and the vector elements are in the interval [-1, 1]. φ represents the search direction parameter. Set the current iteration number as t. The position of the r-th lemming individual in the lemming population is X r (t), and the position of the i-th lemming individual in the lemming population is X i (t). The position of the best lemming individual at the t-th iteration is X′(t). Update the position X i (t) to obtain the position X i (t + 1) of the i-th lemming individual in the lemming population at the (t + 1)-th iteration, where X i (t + 1) = X′(t) + φ·β·(c·(X′(t) - X i (t)) + (1 - c)·(X r (t) - X (t))). The lemming population starts to look for food. Set b1 as a random number in the interval [0, 1], which is the iteration random coefficient. The position of the s-th lemming individual in the lemming population is X s (t). Update the position X i (t + 1) to obtain X i (t + 1) = X i (t) + φ·χ·(X′(t) - X s (t)). At this time, the exploration stage is completed. S222. The lemming population enters the development stage. The lemming population adopts a spiral search strategy, calculates the average distance between the current best lemming individual position and the positions of other lemming individuals, denoted as the search radius R, and sets b2 to represent a random number between the interval [0, 1], the spiral coefficient Use the spiral coefficient to update the position X i (t + 1) to obtain The lemming population evades predators. Levy flight is introduced. The Levy flight function is denoted as Levy(j). The maximum number of iterations is set to T, and the evasion coefficient Update the position X i (t + 1) again to complete the development stage and obtain X i (t + 1) = X′(t) + φ·γ·Levy(j)·(X′(t) - X i (t)), generating the next generation of lemming population; S223. Use the energy factor to balance the exploration stage and the development stage. Set b3 to represent a random number between the interval [0, 1]. Then the energy factor Set the energy threshold. When the energy factor is greater than the energy threshold, the lemming population enters the exploration stage at this time; otherwise, the lemming population enters the development stage. Continuously iterate until the current iteration number reaches the maximum iteration number, then stop the iteration to obtain the final lemming population. Search for the best lemming individual corresponding to the best fitness function value in the final lemming population. The best lemming individual is the global optimal solution, and the global optimal solution includes the optimized engine power, the optimized bus voltage compensation amount, and the optimized battery current, and generate the optimized operating parameters of the hybrid forklift AGV; S23. The hybrid forklift AGV operates according to the optimized operating parameters of the hybrid forklift AGV. At this time, the hybrid forklift AGV realizes preliminary adaptive energy management.

7. A method for intelligent control and power supply of an Internet of Things-based physical interlock, according to claim 6, characterized in that The said S3 includes the following steps: S31. According to the optimized operating parameters of the hybrid forklift AGV and measuring the battery temperature, obtain the optimized operating parameter data matrix; collect the battery current, battery voltage and battery temperature again to form a sample data matrix, train the BP neural network to obtain the BP neural network prediction model, and output the predicted power value. S32. Combine the predicted power value and the working condition label to establish an energy fuzzy control model and output an adaptive electric distribution strategy.

8. A method for intelligent control and power supply of an Internet of Things-based physical lock, according to claim 7, characterized in that The said S32 includes the following steps: S321. Establish a rule base according to the predicted power value and the working condition label. S322. Set the membership function, and the membership function is a normal distribution function. The motor power, driving speed and predicted power value respectively contain several fuzzy subsets. Combine the rule base and the membership function to establish a fuzzy control rule table. According to the change of the driving speed under different working condition labels, input the predicted power value and then output the motor power to complete the energy distribution and obtain the adaptive electric distribution strategy.

9. A method for intelligent control and power supply of an Internet of Things-based physical interlock, according to claim 8, characterized in that The said S4 includes the following steps: S41. The hybrid forklift AGV is respectively in combined drive or motor single drive, and operates according to the optimized operating parameters and the adaptive electric distribution strategy of the hybrid forklift AGV to obtain the adaptive energy distribution strategy, perform energy distribution on the hybrid forklift AGV, and realize adaptive energy management.

10. A system for implementing the adaptive energy management method for a hybrid forklift AGV according to any one of claims 1-9, characterized in that, Specifically including: A working condition label generation module, an operating parameter optimization module, an electric distribution strategy generation module and an adaptive energy management module; The working condition label generation module is used to perform data preprocessing on the initial operating parameter data and then identify the working conditions to generate working condition labels. The operating parameter optimization module is used to solve the optimized operating parameters using the lemming optimization algorithm. The electric distribution strategy generation module is used to establish an energy fuzzy control model and output an adaptive electric distribution strategy. The adaptive energy management module is used to generate an adaptive energy distribution strategy according to the optimized operating parameters and the adaptive electric distribution strategy.

Images